chore: update pattern detection docs

Signed-off-by: Matthew Grigsby <[email protected]>

Author: MatthewGrigsby

docs/reference/llm/index.md

@@ -11,6 +11,7 @@ The **LLM (Large Language Model) Module** provides interfaces and implementation
 
 ### LLM Adapters
 Adapters to support multiple LLM providers:
+
 - **`BaseVendorAdapter`** - Abstract base class for all LLM vendor adapters.
 - **`AnthropicAdapter`** - Adapter for Anthropic’s Claude models.
 - **`MistralAdapter`** - Adapter for Mistral AI models.

docs/reference/llm/pattern_detection/index.md

@@ -8,43 +8,11 @@ The **Pattern Detection Module** provides utilities for detecting predefined pat
 
 ## Core Components
 
-<div style="display: flex; margin-bottom: 20px;">
-<div style="flex: 1; padding-right: 15px;">
-
-### Automaton Classes
-
-* **`AhoCorasickAutomaton`**  
-  ✓ Trie-based pattern matching engine  
-  ✓ Linear-time complexity  
-  ✓ Simultaneous multi-pattern search
-
-* **`AhoCorasickAutomatonNormalized`**  
-  ✓ Whitespace-insensitive matching  
-  ✓ Pattern normalization  
-  ✓ Original-to-normalized index mapping
-
-</div>
-<div style="flex: 1; padding-left: 15px;">
-
-### Processor Classes
-
-* **`BaseBufferedProcessor`**  
-  ✓ Abstract base for text streaming  
-  ✓ Buffer management  
-  ✓ Chunk-based processing
-
-* **`AhoCorasickBufferedProcessor`**  
-  ✓ Exact pattern matching  
-  ✓ YAML-configurable patterns  
-  ✓ Streaming-ready implementation
-
-* **`AhoCorasickBufferedProcessorNormalized`**  
-  ✓ Whitespace-invariant detection  
-  ✓ Flexible text matching  
-  ✓ Preserves original text positions
-
-</div>
-</div>
+| Automaton Classes | Processor Classes |
+|-------------------|-------------------|
+| **`AhoCorasickAutomaton`**<br>✓ Trie-based pattern matching engine<br>✓ Linear-time complexity<br>✓ Simultaneous multi-pattern search | **`BaseBufferedProcessor`**<br>✓ Abstract base for text streaming<br>✓ Buffer management<br>✓ Chunk-based processing |
+| **`AhoCorasickAutomatonNormalized`**<br>✓ Whitespace-insensitive matching<br>✓ Pattern normalization<br>✓ Original-to-normalized index mapping | **`AhoCorasickBufferedProcessor`**<br>✓ Exact pattern matching<br>✓ YAML-configurable patterns<br>✓ Streaming-ready implementation |
+| | **`AhoCorasickBufferedProcessorNormalized`**<br>✓ Whitespace-invariant detection<br>✓ Flexible text matching<br>✓ Preserves original text positions |
 
 ### Utility Functions
 
@@ -58,14 +26,14 @@ The **Pattern Detection Module** provides utilities for detecting predefined pat
 The module processes text in two primary stages:
 
 1. **Pattern Preprocessing**
-   - Patterns are loaded from YAML configuration
-   - The Aho-Corasick automaton is constructed from patterns
-   - Failure links connect states for efficient pattern transitions
+     - Patterns are loaded from YAML configuration
+     - The Aho-Corasick automaton is constructed from patterns
+     - Failure links connect states for efficient pattern transitions
 
 2. **Buffered Text Processing**
-   - Text is processed in manageable chunks
-   - Partial matches at chunk boundaries are preserved
-   - Match information includes pattern name and position
+     - Text is processed in manageable chunks
+     - Partial matches at chunk boundaries are preserved
+     - Match information includes pattern name and position
 
 ### Text Normalization Pipeline
 
@@ -81,16 +49,19 @@ The module processes text in two primary stages:
 
 ## Performance Considerations
 
-* **Time Complexity**: O(n + m + k) where:
+**Time Complexity** O(n + m + k) where:
+
   * n = length of input text
   * m = total length of all patterns
   * k = number of pattern occurrences
 
-* **Space Efficiency**:
+**Space Efficiency**:
+
   * Buffered processing minimizes memory usage
   * Suitable for streaming applications with unbounded input
 
-* **Flexibility vs. Performance**:
+**Flexibility vs. Performance**:
+
   * Standard processors offer exact matching with minimal overhead
   * Normalized processors provide flexibility with slight computational cost